Defect Detection Of Complex Inner Hole Based On Machine Vision

With the rapid development of automobile industry, accessories industry of automobile get a great progress as well. As one of the most important part in the steering system of car, the special of structure of Steering valve set is required to supervise in the aspect of quality inspection strictly. Traditional operation depends on human operator which is too inefficient to ensure inspection quality, and it also causes a huge waste of human resource. In this paper, the writer will discuss the possibilities of using machine vision technology to examine the surface quality of the inner hole in the switching valve.Using the method of machine vision technology, the project has designed to defect defects on the surface of complex inner hole, such as scratch, attrition and gouges intelligently and automatically.Firstly, this paper introduces the research background, significance and research situation at home and broad, then chooses the camera as well as the selection of light source and introduces the structure of the whole system. Secondly, in the processing stage, because some parts of image are complex and not necessary, the system check the edge and corner so that parts which we do not need to examine can be separated. After that, it introduces the algorithm of image spread which do not influence the result of experiment. Spreading images of inner hole is to check parts whose defect has been affirmed by the system through manual operation, which is more aligned with the perspective of mankind. Then, in order to record data characteristic of those defection, we extract feature through algorithm of gray, geometry, texture and dimensionality reduction using algorithm of principal component analysis. Finally, defection is defected by algorithm of Laplacian support vector machine and show the result.